Rationale and Hypothesis: Highly active antiretroviral therapy (HAART) for HIV-1 infection fails to suppress viral replication in 37-70% of children within one year of initiation, due in large part to the selection of drug-resistant mutants. Persistent replication despite therapy, the formula for the selection of drug-resistant viral variants, continues in approximately half of individuals during "effective" HAART (median HIV-1 RNA <50 copies/ml). This ongoing replication has been localized to peripheral blood mononuclear cells (PBMC), circulating monocytes, and tissue macrophages. While inadequate drug levels secondary to difficulty with adherence account for the majority of persistent replication, we hypothesize that pulmonary macrophages, a potential viral reservoir and site of persistent replication, and the host multi-drug resistance genotype (MDR1) C/C genotype, a host genetic factor associated with poor therapeutic outcomes in varied diseases, also have an important role. Understanding contributory factors to persistent viral replication and developing a simple method to detect persistent viral replication will have significant clinical utility. Specific Aims: We propose to (1) determine if drug-resistant HIV-1 is preferentially selected in macrophages; (2) determine if individuals with persistent replication are more likely to have the MDR1 C/C genotype, and if individuals with the MDR1 C/C genotype are more likely to have CXCR4 tropic virus, as well as increased therapy-associated toxicity; and (3) determine if either vRNA or vDNA load is a surrogate marker for persistent replication. Methods: HIV-1 env and pol sequences will be derived from the lung macrophages of 20 children with a mean viral load of <50 copies/ml, including children with low-level or no evidence of viral replication during HAART. The sequences from lung macrophages will be compared to those from blood lymphocytes for evidence of evolution and selection of drug-resistant mutants. Both these children and a large retrospective cohort of children will be genotyped at the MDR1 locus, to determine if those with persistent replication or "virologic failure" have the MDR1 C/C genotype. In addition, sequences obtained from both episodes of low-level viremia and pulmonary macrophages will be assessed for coreceptor usage. Finally, HIV-1 vDNA from pulmonary macrophages and vRNA and vDNA from PBMC populations will be quantified. Conclusion: Genetic analyses will determine if viral replication continues in macrophages when it is suppressed in lymphocyte populations and if subjects with the MDR1 C/C genotype are at higher risk for persistent replication. Levels of vRNA or vDNA in PBMC and vDNA in pulmonary macrophages of the children studied will be evaluated for their utility in predicting which children have low-level viral replication during HAART. Our studies of persistent replication and the selection of drug-resistant mutants could guide clinical trials striving to prolong the efficacy of HAART in treating HIV-1 infected children.